Core Scientific Inc. announced a strategy to expand its Muskogee, Oklahoma campus to approximately 1.5 GW of gross power [1].

The expansion reflects a broader industry shift as data center operators pivot to support the massive energy requirements of artificial intelligence. By scaling its capacity for high-density colocation, the company intends to capture the growing market for AI workloads and digital infrastructure [1].

The multi-tiered strategy was first detailed on May 6, 2024 [1]. The plan focuses on increasing the gross power available at the Muskogee site to 1.5 GW [2]. This scale of power infrastructure is necessary to house the specialized hardware required for modern large-scale computing tasks.

Core Scientific, which trades on the NASDAQ under the symbol CORZ, is positioning the Oklahoma campus as a hub for high-performance computing [3]. The company has targeted the region to leverage available power resources and land for industrial scaling.

While most reports identify the Muskogee, Oklahoma campus as the site of this specific expansion [2], some secondary reports have mentioned other locations. However, the primary strategy focuses on the Oklahoma facility's growth to reach the 1.5 GW threshold [1].

The move comes as the company seeks to diversify its operations beyond traditional cryptocurrency mining. By investing in the infrastructure needed for AI, Core Scientific is attempting to stabilize its revenue streams through long-term colocation contracts with technology firms.

Core Scientific Inc. announced a strategy to expand its Muskogee, Oklahoma campus to approximately 1.5 GW of gross power.

This expansion signals a strategic pivot for Core Scientific, moving from the volatile nature of Bitcoin mining toward the more stable, high-growth sector of AI infrastructure. By securing 1.5 GW of power, the company is building a competitive moat in the 'power-land grab' currently defining the AI era, where access to massive amounts of electricity is the primary bottleneck for scaling large language models.